Monthly Management of Meta and TikTok Paid Social

Meta needs roughly 50 optimisation events per ad set per week to leave the learning phase. Below that threshold, daily performance is noise, and most of what agencies bill as daily optimisation is a reaction to noise. We change one variable at a time, hold it for a full conversion window before judging it, and review structure twice a week. Accounts we take over typically see their weekly change count fall by two thirds in the first month. What we optimise on: Net revenue after returns, pulled from your ERP or Baselinker and joined back to campaign level, not gross platform-reported revenue. In fashion this is the whole argument. A campaign at 4.1 gross ROAS with a 41 percent return rate is losing to one at 3.2 with 18 percent, and an agency reporting only the first number is being paid to look good. We build the returns join in week one or we decline the account. The month, as it actually runs: Two structural reviews a week, Monday and Thursday. Creative goes live in one weekly drop rather than trickling, so read windows stay clean. Budget moves at the campaign level; we do not touch ad set bids to manufacture activity. A monthly written note, four pages at most, that says what we changed, what it did, and what we got wrong. Reporting is delivered in English and Polish where your team needs both. Incrementality: Every retainer above the entry tier carries a standing geo holdout. In Poland we hold out matched voivodeship groups; in export markets, matched country or DMA pairs. Meta Conversion Lift where volume supports it. The point is not the number on the platform dashboard, it is the difference between a market that saw the ads and one that did not. We have run tests that showed no measurable lift and we reported them as such. Creative: We brief and direct. Concept, hook structure, the first two seconds, and what the ad must not say. Production sits with your team or a studio you already use, and we will happily work to the assets you have rather than insisting on a shoot. What we refuse is the volume treadmill: forty variants a month of the same idea is a way to bill hours, not a way to find a winner. What we refuse: Performance-based pricing, because it depends on attribution neither of us controls and it quietly pushes an agency toward branded search and retargeting. Last-click reporting. Managing an account whose event data we have not verified, which is a separate engagement and comes first. Running a channel we do not believe belongs in your mix so the retainer looks fuller. What sits outside scope: Google, retail media and Amazon. Email and CRM. Landing page development, though we will tell you plainly when the page is the constraint and stop recommending budget increases until it changes. Influencer contracting, which is its own engagement. Who this is not for: Budgets below roughly 60,000 EUR a month in working media, where a fee this size distorts the account. Businesses that need spend live next week; onboarding is three to four weeks and the first two are measurement. Teams who want a partner available in a shared channel within minutes. We answer within a working day, and we would rather be judged on the Thursday review than on response time. Who this suits: In-house leads who already have a functioning creative pipeline and want the media decisions made by someone senior enough to say the quarter is not working.